Spaces:
Sleeping
Sleeping
File size: 1,357 Bytes
8293818 14632bf 8293818 2cc4c0e 8293818 2cc4c0e 8293818 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 |
import gradio as gr
import tensorflow as tf
import numpy as np
from PIL import Image
# Load the trained model
model = tf.keras.models.load_model('tato.h5')
# Define class labels (update with your dataset's class names)
class_labels = ['Late Blight', 'Early Blight', 'Healthy']
# Define a prediction function
def predict(image):
# Resize and preprocess the image
image = image.resize((224, 224)) # Resize to match model input size
image_array = np.array(image) / 255.0 # Normalize the image
image_array = np.expand_dims(image_array, axis=0) # Add batch dimension
# Make predictions
predictions = model.predict(image_array)
predicted_class = class_labels[np.argmax(predictions)] # Map prediction to class label
confidence = np.max(predictions) # Get the highest confidence score
return f"Predicted Class: {predicted_class}" #with confidence {confidence:.2f}"
# Create a Gradio interface
interface = gr.Interface(
theme="Subh775/orchid_candy",
fn=predict, # The prediction function
inputs=gr.Image(type="pil"), # Input type (image as PIL object)
outputs="text", # Output type (text)
title="Plant Disease Classifier",
description="Upload an image of a potato plant leaf to identify its condition (Early_blight, Late_blight, Healthy)"
)
# Launch the interface
interface.launch()
|